louis_lifu
/
vllm
mirror of https://github.com/vllm-project/vllm
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Use slow tokenizer for open llama models (#168)

This commit is contained in:
Woosuk Kwon 2023-06-19 23:19:47 -07:00 committed by GitHub
parent 033f5c78f5
commit 67d96c29fb
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
vllm/engine/tokenizer_utils.py
View File

@ -17,7 +17,12 @@ def get_tokenizer(
) -> Union[PreTrainedTokenizer, PreTrainedTokenizerFast]:
"""Gets a tokenizer for the given model name via Huggingface."""
config = AutoConfig.from_pretrained(model_name)
if config.model_type == "llama" and getattr(kwargs, "use_fast", True):
if "open_llama" in model_name:
kwargs["use_fast"] = False
logger.info(
"OpenLLaMA models do not support the fast tokenizer. "
"Using the slow tokenizer instead.")
elif config.model_type == "llama" and getattr(kwargs, "use_fast", True):
# LLaMA fast tokenizer causes protobuf errors in some environments.
# However, we found that the below LLaMA fast tokenizer works well in
# most environments.